WO2014037283A1

WO2014037283A1 - Device and method for supplying a reference audio signal to an acoustic processing unit

Info

Publication number: WO2014037283A1
Application number: PCT/EP2013/067960
Authority: WO
Inventors: Etienne EVRARD
Original assignee: Sagemcom Broadband Sas
Priority date: 2012-09-06
Filing date: 2013-08-30
Publication date: 2014-03-13
Also published as: BR112015004905A2; FR2995122A1; US9412378B2; FR2995122B1; CN104769670B; CN104769670A; EP2893531B1; US20150228282A1; BR112015004905B1; EP2893531A1

Abstract

The invention relates to an apparatus comprising a first interface intended to be connected to a sound reproduction device and at least one second interface intended to be connected to at least one microphone, and an acoustic processing unit suitable for outputting an audio signal that is filtered by attenuating or suppressing a reference audio signal of an audio signal received from the second interface. In an initialization stage, the apparatus determines (405) propagation delay, configures (406) a buffer memory using a reading trigger threshold defined on the basis of the determined propagation delay. In a nominal operating stage, the apparatus transmits a third audio signal via the first interface, the third audio signal being the reference signal after the latter has passed through the memory buffer.

Description

Device and method for providing a reference audio signal to an acoustic processing unit

The present invention relates to an equipment comprising a first interface intended to be connected to a sound reproduction device and at least one second interface intended to be connected to at least one microphone, said equipment comprising an acoustic processing unit adapted to deliver an audio signal filtered by attenuating or deleting a reference audio signal from an audio signal received via said second interface.

Today there are many applications for voice recognition and thus to control devices by voice. The difficulty lies in being able to distinguish these voice commands from a noisy sound environment.

We find the same type of problem in the world of teleconferencing. It can sometimes be difficult to clearly distinguish the words of an interlocutor because of a noisy sound environment.

This noisy sound environment is often linked to audio or audiovisual content that is played back during the conference call or while the user is transmitting voice control. For example, the case of a home-based system cinema ("home-theater" in English) that we would like to order by voice. When the home theater system is turned on, an audible signal is played back in the room, potentially with a loud volume. It would be difficult for the system to distinguish voice commands in this situation.

There are components for removing a reference audio signal from an audio signal picked up by one or more microphones. Electronic evaluation boards are available on the shelf based on such components. However, these components and evaluation boards do not provide a satisfactory result in many installation configurations. Indeed, if we take the example of the aforementioned home-cinema system, the propagation time in the air of sound signals coming from loudspeakers depends on the actual location of these loudspeakers and the configuration of these Electronic components and boards are often unsuitable, so the component can not find the reference signal in the audio signal picked up by the microphone.

It is therefore desirable to overcome these disadvantages of the state of the art.

The invention relates to an equipment comprising a first interface intended to be connected to a sound reproduction device and at least one second interface intended to be connected to at least one microphone, said equipment comprising an acoustic processing unit adapted to deliver an audio signal filtered by attenuating or deleting a reference audio signal from an audio signal received via said second interface. The equipment is such that it comprises means for implementing an initialization phase, comprising: means for determining a propagation latency between a moment of transmission of a first audio signal via said first interface and an instant of receiving a second audio signal via said second interface; means for configuring a buffer with a read trigger threshold defined according to said determined propagation latency. The equipment is such that it further comprises means for implementing a nominal operating phase, comprising: means for transmitting a third audio signal via said first interface, said third audio signal being said reference signal after passing through said buffer memory.

Thus, it is possible to perfectly adapt the configuration of the equipment to various situations in which the voice of a user must be distinguished in a noisy or even very noisy sound environment because of an audio signal that the equipment sends to the device of restitution. According to a particular embodiment, said first audio signal consists of a predefined pattern.

According to a particular embodiment, said means for determining the propagation latency comprise means for detecting an amplitude threshold crossing of said second audio signal.

According to a particular embodiment, said means for determining the propagation latency comprise a North filter applied between said first audio signal and said second audio signal.

According to a particular embodiment, said means for implementing the initialization phase are implemented in a control unit connected to an output of the acoustic processing unit for receiving said filtered audio signal, and the control unit deactivates any reference audio signal transmission to said acoustic processing unit during said initialization phase.

According to a particular embodiment, said means for implementing the initialization phase, said first interface and said means for transmitting the third audio signal are implemented in a first device, the acoustic processing unit and said second interface are implemented. in a second device intended to be connected to said first device.

According to a particular embodiment, the said microphone (s) being unidirectional, the acoustic processing unit and the said microphone (s) are implemented in a housing having, for each microphone, a first slot and a second slot, and each microphone is installed in a cavity of a holder in which a first slot and a second slot are also made, and placed in respective correspondence with said slots of the case when said support is mounted in said housing, said support being adapted so that the distances between said first slots and a face of said microphone placed in the direction of a preferred sound signal and the distances between said second slots and an opposite face of said microphone are substantially identical.

The invention also relates to a method implemented by a device comprising a first interface intended to be connected to a sound reproduction device and at least one second interface intended to be connected to at least one microphone, said equipment comprising a processing unit. acoustic device adapted to deliver a filtered audio signal by attenuation or deletion of a reference audio signal from an audio signal received via said second interface. The method is such that it comprises an initialization phase comprising the following steps: determining a propagation latency between a moment of transmission of a first audio signal via said first interface and a moment of reception of a second audio signal via said second interface; configuring a buffer with a read trigger threshold defined according to said determined propagation latency. The method is such that it further comprises a nominal operating phase comprising the step of: transmitting a third audio signal via said first interface, said third audio signal being said reference signal after passing through said buffer memory.

The invention also relates to a computer program, which can be stored on a medium and / or downloaded from a communication network, in order to be read by a processor. This computer program includes instructions for implementing any of the methods mentioned above, when said program is executed by the processor. The invention also relates to storage means comprising such a computer program.

The characteristics of the invention mentioned above, as well as others, will appear more clearly on reading the following description of an exemplary embodiment, said description being given in relation to the attached drawings, among which:

FIG. 1 schematically illustrates a system in which the invention can be implemented;

FIG. 2A schematically illustrates an example of a hardware architecture of a source device of the system of FIG. 1;

FIG. 2B schematically illustrates an example of a hardware architecture of an acoustic processing device of the system of FIG. 1;

FIG. 3 schematically illustrates another example of hardware architecture of the source device;

FIG. 4 schematically illustrates an initialization algorithm of the source device;

FIG. 5 schematically illustrates a nominal operating algorithm of the source device;

FIG. 6 schematically illustrates a simplified perspective view of a shell of a housing in which the acoustic treatment device can be installed; FIG. 7A schematically illustrates a perspective view of a microphone holder for placement in the housing;

FIG. 7B schematically illustrates another view of the microphone holder. In a system in which a rendering device is intended to reproduce, in the form of a sound signal, an audio signal supplied by a source equipment, it is proposed to implement an initialization phase in which a propagation latency is determined between the transmission of the audio signal by the source equipment and the reception of a corresponding sound signal by at least one microphone intended to capture at least the voice of a user in a nominal operating phase. A buffer is then configured with a read trigger threshold defined according to the determined propagation latency. Then, in the nominal operating phase, when the equipment transmits an audio signal to the playback device, the equipment also transmits it to the buffer which causes a delay. An acoustic processing unit adapted to deliver a filtered audio signal by attenuation or deletion of a reference audio signal from a received audio signal is used, with input that captures the microphone (s). ) and as a reference signal, the signal delayed by the buffer memory.

Fig. 1 schematically illustrates a system in which the invention can be implemented.

The system of FIG. 1 comprises a source device 103 of audio or audiovisual signal. According to a first example, the source device 103 is a digital decoder adapted to receive and decode audiovisual signals from a satellite link or an Ethernet link to a home gateway via which audiovisual content is received from the Internet. According to a second example, the source device 103 is a Blu-Ray (registered trademark) player or a computer on which a media player software ("media player") runs. Any device adapted to provide an audio signal to be reproduced by a sound reproduction device can be implemented.

The system of FIG. 1 further comprises a sound reproduction device 101, which may be an audiovisual playback device. According to a first example, the sound reproduction device 101 is a screen comprising integrated speakers. In a second example, the sound reproduction device 101 is a hi-fi amplifier. The source device 103 comprises an interface 151 to be connected to the sound reproduction device 101 by means of a link 141. The sound reproduction device 101 includes an interface 110 to be connected to the source device 103 via the link 141. For example , the link 141 complies with the HDMI ("High-Definition Multimedia Interface"), WHDI ("Wireless Home Digital Interface"), SPDIF ("Sony / Philips Digital Interconnect Format") or Scart (brand Mark). Thus, the sound reproduction device 101 is capable of reproducing any audio signal received from the source device 103 via the link 141.

The system of FIG. 1 further comprises an acoustic treatment device

102, and at least one microphone 111, 112. The acoustic processing device 102 comprises at least one interface 121, 122, adapted to connect the microphone (s) 111, 112. The acoustic processing device 102 is thus suitable receiving audio signals corresponding to sound signals picked up by the microphone (s) 111, 112. The acoustic processing device 102 also has an interface 123 to be connected to the source device 103 via a link 142 The source device 103 includes an interface 153 for connection to the sound processing device 102 via the link 142. For example, the link 142 conforms to the HDMI, USB ("Universal Serial Bus") or IEEE 1394 specifications.

The microphone (s) 111, 112 can capture a sound environment, and in particular the sound signals broadcast by the playback device 101 and the voice of a user of the system.

The source device 103 and the acoustic processing device 102 may be incorporated in the same housing, and may also be implemented on the same printed circuit board (PCB for "Printed Circuit Board" in English), the link 142 then being a circuit board track.

Fig. 2A schematically illustrates an example of hardware architecture of the source device 103.

The source device 103 comprises an audio signal supply unit 211 for providing an audio signal, for example resulting from a demultiplexing and decoding of an audiovisual signal received via a satellite link. The audio signal is supplied to the interface 151 and to the input of a FIFO buffer ("First-In First-Out") 202 of the source device 103. initialization, the audio signal is also supplied to a control unit 203 of the source device 103. During the initialization phase, another audio signal, coming from the interface 153, is also supplied to the control unit 203.

The sound signal supply unit 211 may further comprise an audio signal generator according to a predefined pattern, which can be used during the initialization phase.

The aforementioned initialization phase is detailed below in relation to FIG. 4, and a subsequent nominal operation phase of the source device 103 is detailed hereinafter with reference to FIG. 5.

The source device 103 includes a processing unit 212 for applying a processing on a filtered audio signal from the interface 153. According to a first example, the processing unit 212 implements a voice recognition mechanism. In a second example, the processing unit 212 implements a shaping mechanism for transmitting the filtered audio signal in a teleconference.

Fig. 2B schematically illustrates an example of hardware architecture of the acoustic processing device 102.

The acoustic processing device 102 comprises an acoustic processing unit 201 whose function is to delete a first audio signal, referred to as a reference signal, from a second audio signal. The reference audio signal is provided by the source device via the link 142. The second audio signal is the audio signal resulting from the sound signal picked up by the microphone (s) 111, 112. The acoustic processing unit 201 provides then to the source device 103 via the link 142 a filtered audio signal, i.e. as far as possible devoid of the reference audio signal, when this reference audio signal has been detected in the signal picked up by the microphone (s) (s). For example, the acoustic processing unit 201 is a CX20708-21X reference component of the Connexant company.

It should be noted that the acoustic processing unit 201 may comprise an internal buffer memory to enable processing to be performed by the acoustic processing unit 201. However, this internal buffer memory serves only to store the audio signals. during a predefined time window, for example of the order of 200 ms, to perform these treatments. No read trigger is associated with it and can not be configured. Figs. 2A and 2B are respectively complementary examples of hardware architecture of the source device 103 and the acoustic processing device 102. A different distribution of the functions implemented can be envisaged. For example, the control unit 203 and / or the FIFO 202 can be implemented in the acoustic processing device 102. The arrangement according to Figs. 2A and 2B, however, has the advantage of allowing an easy update of devices providers of audio or audiovisual content already deployed, for example at home. Indeed, if we take the example of the satellite decoders or television over IP ("Internet Procotol" in English), they implement many functions by software implementation. It is then easy to update this software implementation in order to implement the functions described here in relation to the source device 103. It would then suffice to add the acoustic processing device 102 to implement the invention. , without having to replace the hardware platform of these decoders.

The term "equipment" will be used to designate indifferently a device or a set of devices implementing these functions.

Fig. 3 schematically illustrates another example of hardware architecture of the source device 103, which then comprises, connected by a communication bus 310: a processor or CPU ("Central Processing Unit" in English) 300; Random Access Memory (RAM) 301; ROM (Read Only Memory) 302; a storage unit or a storage medium reader, such as a hard disk drive HDD ("Hard Disk Drive") 303; a first interface 304 for communicating via the link 141; and a second interface 305 for communicating via the link 142.

It should be noted that the acoustic processing device 102 can be implemented with a similar hardware architecture.

In the context of the architecture presented in FIG. 3, the implementation of the FIFO 202 can be done within the second interface 305 or within the RAM 301, for example in the form of a linked list.

The processor 300 is capable of executing loaded instructions in the RAM

301 from the ROM 302, an external memory (not shown), a storage medium, such as the HDD hard disk 303, or a communication network. When the source device 103 is turned on, the processor 300 is able to read instructions from RAM 301 and execute them. These instructions form a computer program causing the implementation, by the processor 300, of all or part of the algorithms and steps described below. All or part of the algorithms and steps described below can be implemented in software form by executing a set of instructions by a programmable machine, such as a DSP ("Digital Signal Processor" in English) or a microcontroller, or be implemented in hardware form by a machine or a dedicated component, such as an FPGA ("Field Programmable Gate Array" in English) or an ASIC ("Application-Specific Integrated Circuit").

Fig. 4 schematically illustrates an algorithm implementing an initialization phase 400 of the source device 103.

In a step 401, the source device 103 sends an audio signal via the interface 151. This audio signal preferably corresponds to a predefined pattern. This audio signal can also be a priori unknown to the source device 103, for example from a demultiplexing and decoding of an audiovisual signal received by the source device 103 via a satellite link. In the architecture shown in FIG. 2A, the source device 103 also sends the audio signal to the control unit 203.

In a subsequent step 402, the source device 103 determines information representative of an instant at which the source device 103 has sent the audio signal via the interface 151. The audio signal sent by the source device 103 is therefore intended to be reproduced by the playback device 101. The playback device 101 decodes the audio signal transmitted by the source device 103 and generates a corresponding sound signal, the microphone (s) 111, 112 being adapted (s) to capture this sound signal.

In a subsequent step 403, the source device 103 performs, or requires, a sound environment capture. To do this, the source device 103 instructs, via the link 142, the acoustic processing device 102 to start a sound environment capture using the microphone (s) 111, 112. During this initialization phase, the source device 103 does not transmit a reference audio signal to the acoustic processing device 102 via the interface 153. The acoustic processing device 102 then transmits back to the source device 103 directly the audio signal corresponding to the sound signal captured by the microphone (s). ) 111, 112, no reference audio signal to be removed from the sound signal picked up by the microphone (s) 111, 112. In the architecture shown in FIG. 2 A, the device source 103 sends the audio signal received from the acoustic processing device 102 to the control unit 203.

In a next step 404, the source device 103 determines information representative of an instant at which the acoustic processing device 102 has received the audio signal through the microphone (s) 111, 112. It is possible to consider that the the instant at which the acoustic processing device 102 has received the audio signal through the microphone (s) 111, 112 is the same as that to which the source device 103 has received the audio signal via the interface 153. It is then considered that the treatments performed by the acoustic treatment device 102 have a negligible latency. If this is not the case, the source device 103 has, by configuration, knowledge of this latency time and can thus take it into account.

In order to determine the instant at which the acoustic processing device 102 has received the audio signal through the microphone (s) 111, 112, the source device 103 detects a time when the audio signal received via the interface 153 is greater than at a predefined threshold. The source device 103 then considers that this moment of crossing of said predefined threshold is the one at which the acoustic processing device 102 has received the audio signal thanks to the microphone (s) 111, 112. According to an alternative embodiment, the source device 103 performs a correlation between the audio signal sent via the interface 151 and the audio signal received via the interface 153, in order to determine which time window corresponds to the audio signal sent in the received audio signal. To do this, a matching filter ("matched filter" in English), also called North filter, can be applied. The use of such a filter advantageously makes it possible to maximize the signal-to-noise ratio. Other correlation methods can be used.

In a next step 405, the source device 103 determines information representative of propagation latency, which is the difference between the instant at which the acoustic processing device 102 received the audio signal through the microphone (s). 111, 112 and the time at which the source device 103 sent the audio signal via the interface 151. This propagation latency is determined by the information determined in steps 404 and 402 respectively.

In a next step 406, the source device 103 determines information representative of a FIFO trip threshold 202, to be implemented during the nominal operating phase of the source device 103. This FIFO trip threshold 202 is determined. according to the propagation latency determined in step 405, and makes it possible to apply a delay to the reference audio signal to be transmitted to the acoustic processing device 102 via the interface 153. If the propagation time between the source device 103 and the acoustic processing device 102 is neglected, this delay is equal to the propagation latency determined in step 405. Otherwise, this delay is equal to the propagation latency determined in step 405, at which a predefined value of the propagation time between the source device 103 and the acoustic processing device 102 is subtracted.

Then, the source device 103 configures the FIFO 202 so that the triggering threshold determined in step 405 is applied in the nominal operating phase. The initialization phase and the nominal operating phase are then terminated. can begin.

The source device 103 may provide the user with an indication that the initialization phase is in progress, for example by means of a LED ("Light-Emitting Diode") of a user interface. This may allow the user to know that he must limit any ambient noise to facilitate detection of the audio signal expected back by the source device 103.

In the architecture shown in FIG. 2A, steps 402, 404, 405 and 406 are performed by the control unit 203.

Fig. 5 schematically illustrates a nominal operating algorithm of the source device 103, after the initialization phase has been executed.

In a step 501, the source device 103 activates the filling of the FIFO

202. No data is then present in FIFO 202.

In a next step 502, the source device 103 activates the sound signal capture by means of the microphone (s) 111, 112. To do this, the source device 103 sends to the acoustic treatment device 102 an instruction to engage a such capture. An audio signal corresponding to the sound signal captured by the microphone (s) 111, 112 is then received by the acoustic processing unit 201.

In a subsequent step 503, the source device 103 activates the sending of an audio signal to the reproduction device 101 via the interface 151. This audio signal results, for example, from a demultiplexing and a decoding of an audiovisual content. received or read by the source device 103. The source device 103 having activated the filling of the FIFO 202, the audio signal is also stored in the FIFO 202.

In a next step 504, the source device 103 checks whether the filling threshold of the FIFO 202, determined in the step 406, is reached. If this is the case, a step 505 is performed; otherwise, step 504 is repeated. In step 505, the source device 103 activates the reading of the FIFO 202. The data stored in the FIFO 202 is then transmitted as a reference audio signal to the acoustic processing device 102 via the interface 153. This read the FIFO 202 is performed at the rate where the data of the audio signal is written in the FIFO 202. A time delay, whose duration is adapted to compensate for the propagation latency determined in the step 405, is thus applied to the audio signal provided by the source device 103 to the acoustic processing device 102.

Thus, by applying this delay, the audio signals input to the acoustic processing unit 201 are sufficiently synchronized to allow the acoustic processing unit 201 to remove the reference audio signal from the audio signal corresponding to the signal sound captured by the microphone (s) 111, 112. In this way, the audio signal supplied to the processing unit 212 is filtered and substantially devoid of the sound signal corresponding to the audio signal returned by the playback device 101. A slight noise may, however, remain due to the distortions in the sound signal picked up by the microphone (s) 111, 112 relative to the reference audio signal. Then, when the user wishes to use voice commands or participate in a teleconference, his voice can be clearly distinguished in the audio signal, even if the sound volume of the playback device 101 is high.

Fig. 6 schematically illustrates a simplified perspective view of a shell of a housing 600 in which the acoustic processing device 102 may be installed.

The shell of the casing 600 comprises a first part 601 and a second part 602. The two parts 601 and 602 are intended to be assembled with each other, for example by gluing, or by means of assembly screws, or using clips.

Preferably, said first portion 601 serves as a cover for said second portion 602. The outer thickness of this first portion is shown in phantom in FIG. 6. The acoustic treatment device 102 consists of an electronic card on which are mounted components performing the previously described functions. The electronic card is mounted on the inner face of said first part 601. The assembly of the electronic card with said first part 601 can be done by means of assembly screws, rivets or clips. The microphones 111, 112 are furthermore integrated into the casing 600, the shell of which comprises, for each microphone 111, 112, a first slot 610 and a second slot 611. These slots 610, 611 enable the microphones 111, 112 to pick up the microphone. sound environment, as described below in relation to FIGS. 7A and 7B. In the illustration of FIG. 6, the slots 610, 611 are formed in said second portion 602 of the shell.

Fig. 7A schematically illustrates a perspective view of a microphone holder 701 intended to be placed in the housing 600. Each microphone 111, 112 is then unidirectional and has a support 701 associated.

The support 701 intended to receive, in a controlled manner, the microphone 111 or 112 is preferably made of rubber, so as to isolate the microphone 111 or 112 from vibrations transmitted by mechanical parts of the casing 600. Reference microphones CM1045RFH-35BL- C56F1K-LF from the company MWM Acoustics are for example used.

On one side of the support 701 are made two slots 710, 711, intended to be respectively placed in correspondence of the slots 610, 611 formed in the shell of the housing 600, when the support 701 is installed in the housing 600.

The support 701 has a cavity 702 opening on the slots 710, 711, and intended to receive the microphone 111 or 112. Once installed in the cavity 702, the microphone 111 or 112 is arranged so that one side of the microphone 111 or 112 in the direction of the sound signal to be preferred is in the extension of the slot 711 and therefore of the slot 611; in addition, the microphone 111 or 112 is arranged so that the face of the microphone 111 or 112 opposite to the direction of the sound signal to be preferred is in the extension of the slot 710 and therefore of the slot 610. In the typical constructions of the unidirectional microphones, this opposite face to the direction of the sound signal to be preferred has a hole, which allows to let in the background sounds other than those to be preferred, that is to say those that come from other directions than the one where does the sound signal come from? In other words, this hole reduces the ambient noise, without eliminating it. To do this, however, it is necessary that the propagation times of the sound signal from the slots in the shell of the housing 600 and the two aforementioned faces of the microphone 11, 112 are substantially identical, that is to say that the distances between these slots and these faces are substantially identical. The arrangement of the support 701 makes it possible to achieve this objective. The term "Substantially" means that any existing difference is negligible in view of the responsiveness of the microphone.

The combination of the support 701 as presented and at least one microphone 111, 112, thus makes it possible to highlight the sound signal (voice of the user in the system of Fig. 1) in the direction preferred by the s) microphone (s) 111, 112 unidirectional (s). "Highlighting" means bringing out the user's voice compared to other sounds in the sound environment. This facilitates the processing performed by the acoustic processing device 102.

Fig. 7B schematically illustrates another view of the microphone holder 701. It appears, more strikingly in this view, that the slot 71 1 is preferably formed by a recess, or recess. The slot 711 is thus definitively formed when the support 701 is mounted bearing on the internal face of the first part 601 of the shell of the casing 600.

It should be noted that the support 701 can be used to install a unidirectional microphone in a housing without implying the processing implemented by the source devices 103 and acoustic treatment 102. This improves the focus of the voice of the user.

Claims

1) Equipment (102, 103) having a first interface (151) for connection to a sound reproduction device and at least one second interface (121; 122) for connection to at least one microphone (111; 112) , said equipment comprising an acoustic processing unit (201) adapted to deliver an audio signal filtered by attenuation or deletion of a reference audio signal from an audio signal received via said second interface, characterized in that it comprises means for implementing an initialization phase comprising:

means for determining (405) a propagation latency between a transmission instant of a first audio signal via said first interface and a reception instant of a second audio signal via said second interface;

means for configuring (406) a buffer memory (202) with a read trigger threshold defined according to said determined propagation latency;

and in that it further comprises means for implementing a nominal operating phase comprising:

means for transmitting (503) a third audio signal via said first interface, said third audio signal being said reference signal after passing through said buffer.

2) Equipment according to claim 1, characterized in that said first audio signal consists of a predefined pattern. 3) Equipment according to any one of claims 1 and 2, characterized in that said means for determining the propagation latency comprise means for detecting an amplitude threshold crossing of said second audio signal. 4) Equipment according to any one of claims 1 and 2, characterized in that said means for determining the propagation latency comprises a North filter applied between said first audio signal and said second audio signal. 5) Equipment according to any one of claims 1 to 4, characterized in that said means for implementing the initialization phase are implemented in a control unit connected to an output of the acoustic processing unit to receive said filtered audio signal, and in that the control unit deactivates any reference audio signal transmission to said acoustic processing unit during said initialization phase.

6) Equipment according to any one of claims 1 to 5, characterized in that said means for implementing the initialization phase, said first interface and said means for transmitting the third audio signal are implemented in a first device (103). )

in that the acoustic processing unit and said second interface are implemented in a second device (102) to be connected to said first device.

7) Equipment according to any one of claims 1 to 6, characterized in that the (s) said (s) microphone (s) being unidirectional (s), the acoustic treatment unit and the (s) said (s) ) microphone (s) are implemented in a housing having, for each microphone, a first slot (610) and a second slot (611),

and in that each microphone is installed in a cavity (702) of a support

(701) wherein a first slot (710) and a second slot (711) are also made, and placed in respective correspondence with said housing slots when said holder is mounted in said housing,

said support being adapted so that the distances between said first slots and a face of said microphone placed in the direction of a preferred sound signal and the distances between said second slots and an opposite face of said microphone are substantially identical.

8) A method implemented by a device (102, 103) having a first interface (151) to be connected to a sound reproduction device and at least a second interface (121; 122) for connection to at least one microphone (111; 112), said equipment comprising an acoustic processing unit (201) adapted to deliver an audio signal filtered by attenuation or deletion of a reference audio signal from an audio signal received via said second interface, characterized in that it comprises an initialization phase comprising the following steps:

determining (405) a propagation latency between a transmission time of a first audio signal via said first interface and a timing of reception of a second audio signal via said second interface;

configuring (406) a buffer (202) with a read trigger threshold defined according to said determined propagation latency;

and in that it further comprises a nominal operating phase comprising:

9) Computer program product, characterized in that it comprises instructions for implementing, by an equipment, the method of claim 8, when said program is executed by a processor of said equipment.

10) Storage means, characterized in that they store a computer program comprising instructions for implementing, by an equipment, the method according to claim 8, when said program is executed by a processor of said equipment.